What question did this study set out to answer?

To validate the diagnostic performance of the PM50-B oximetry device against reference sleep studies for obstructive sleep apnea.

February 20, 2026Open Access

Novel Type IV Monitoring Device With Advanced Oximetry Indicators Offers Accurate Diagnosis of Obstructive Sleep Apnea in Adults

Key Points

To validate the diagnostic performance of the PM50-B oximetry device against reference sleep studies for obstructive sleep apnea.
Conducted a prospective observational study.
Analyzed data from 475 adults with suspected obstructive sleep apnea.
Used the PM50-B in conjunction with in-laboratory polysomnography and home sleep apnea testing.
Applied a standardized workflow for signal processing and hypoxemia metric derivation.
ODI2.5_5 showed moderate-to-good agreement with apnea–hypopnea index (ICC = 0.710).
The device demonstrated good discrimination for moderate-to-severe OSA (AUC = 0.925).
Consistency in diagnostic performance across reference modalities (AUC range 0.928–0.983).
Sensitivity and specificity were found to be 81.20% and 91.00%, respectively.

Abstract

ABSTRACT Background Type IV sleep monitors offer a low‐burden option for obstructive sleep apnea (OSA) screening, yet their accuracy is often limited by motion artifacts and variability in signal‐processing methods. The PM50‐B is a wrist‐worn Type IV device that combines high‐sampling‐rate (200 Hz) photoplethysmography (PPG) ‐based oximetry with wrist actigraphy to reduce motion artifacts and employs an adaptive SpO 2 waveform‐based desaturation detection algorithm. This study aimed to validate the diagnostic performance of the PM50‐B against reference sleep studies. Methods In this prospective observational study, adults with suspected OSA underwent simultaneous overnight recording with the PM50‐B and a reference test: in‐laboratory polysomnography (Type I), unattended polysomnography (Type II), or Type III home sleep apnea testing (HSAT). Oximetry and actigraphy signals were processed using a standardized workflow incorporating motion‐artifact attenuation, signal stabilization, and sleep–wake estimation. From the SpO 2 signal, hypoxemia metrics were derived, including the oxygen desaturation index ODI2. 5₅ (≥ 2. 5% desaturation lasting ≥ 5 s/h of total sleep time), cumulative time with SpO 2 < 90% and < 95% (CT90, CT95), and lowest SpO 2. Agreement with the reference apnea–hypopnea index (AHI) was assessed using intraclass correlation coefficients, and diagnostic accuracy was evaluated at clinically relevant AHI thresholds. Results A total of 475 participants were analysed (Type I, n = 37; Type II, n = 32; Type III, n = 406). ODI2. 5₅ showed moderate‐to‐good agreement with AHI (ICC = 0. 710) and good discrimination for moderate‐to‐severe OSA (AHI ≥ 15 events/h), with an under the curve (AUC) of 0. 925 (sensitivity 81. 20%, specificity 91. 00%). Diagnostic performance was consistent across reference modalities (AUC range, 0. 928–0. 983). Conclusion The PM50‐B provides clinically acceptable accuracy for OSA screening when combined with a standardized signal‐processing approach, particularly in comparison with Type III HSAT. ODI2. 5₅ emerged as the strongest diagnostic metric, while CT90, CT95, and lowest SpO 2 provided complementary characterization of nocturnal hypoxemia.

Read Full Paperexternally

Ask AI

Mark Helpful

Bookmark

Relay

View Full Paper